WEL-DINCS RCSEARCH SUPPLEMENT TO THE WELDING JOURNAL, NOVEMBER 1976

Sponsored by the American Welding Society and the Welding Research Council

Material Variables Affecting Lamellar Tearing Susceptibility in Steels

Detailed investigation shows the tendency to lamellar tearing is too complex a phenomenon to fit the more commonly accepted susceptibility criteria

BY S. GANESH AND R. D. STOUT

ABSTRACT. A quantitative weldabi l­ity test was employed to determine the lamellar tearing susceptibil ity of steels with a variety of melting and deoxidation practices, composit ions and thicknesses. The degree was ex­amined to which susceptibil ity was related to inclusion type and dis­t r ibu t ion , oxygen content , m ic ro-st ructura l features, and f rac ture mechanisms. While tearing was gen­erally associated with the elongated nonmetall ic inclusions, in some steels initiation occurred by splitt ing along ferrite bands, l iquation or intergran­ular cracking. Susceptibil ity was ob­served to increase with an increase in the oxygen content of the steel, but the inclusion count obtained using the Quantimet 360 did not show a satisfactory correlation. Voids formed either by decoherence or cracking of inclusions, and terrace l inkage of ad­jacent voids occurred by several modes such as necking, microvoid

S. GANESH is with the Bendix Research Laboratories, Southfield, Michigan, and Ft. D. STOUT is Dean, Graduate School, Lehigh University, Bethlehem, Penn­sylvania.

The work was sponsored by the Welding Research Council and was presented at the 57th AWS Annual Meet­ing held at St. Louis, Missouri, May 10-14, 1976.

coalescence, quas i -c leavage and intergranular cracking. Susceptibil ity was dependent on the nature of void format ion, extent of void growth and the mode of void l inkage. Tech­niques are listed for improving mate­rial resistance to lamellar tearing.

Introduct ion

Lamellar tearing, a form of crack­ing occurr ing in planes essentially parallel to the rolled surface of a plate u n d e r h i g h t h r o u g h - t h i c k n e s s loading, tends to initiate by the deco­herence or cracking of elongated in­clusions. Voids form which grow and link together by the plastic tearing of the intervening matrix, along the horizontal and the vertical planes, producing a characteristic step-l ike appearance to the fracture. Though welding is not a necessary condit ion, lamellar tearing has been generally associated with welded joints and oc­curs in the base metal with insuffi­cient short-transverse ductil ity when subjected to high through-thickness strains generated if weld thermal con­traction is inhibited by structural re­straint.

Tear ing typical ly occurs in the multipass weldments of T, corner, or cruci form joints and is reported in bui ld ing cons t ruc t ion (Refs. 1-4), bridge girders (Ref. 5), pressure-vessels (Ref. 6), shipbui lding (Ref. 7),

Space for UTS Testing

i , \ .i Restraint "~~^ Block

Fig. 1 — Lehigh lamellar tearing fixture

off-shore structures (Ref. 8), boilers (Refs. 8, 10) and nuclear power plants (Ref. 11). Though, in most cases, the r isk of t ea r i ng can be avo ided through proper modif ications to joint design and welding condit ions (Refs. 12, 13), there are situations where the only choice is to select a costlier material having a high resistance to tearing.

Thus, the problem of lamellar tear­ing has aroused much concern (Refs. 1-4, 14) among des igners, steel manufacturers and fabricators; and research efforts have been aimed towards developing weldability tests to assess material susceptibil ity to

WELDING RESEARCH S U P P L E M E N T I 341-S

lamellar tearing and to correlate the susceptibil i ty to short-transverse ten­sile and notch ductility, inclusion dis­tr ibution, and ultrasonic measure­ments.

Despite numerous investigations, limited agreement has emerged on the factors responsible for tearing. There are confl icting opinions re­garding the influence of factors such as melting and deoxidation practice (Refs. 7, 15, 16). plate thickness (Refs. 15, 17-19), rolling direction (Refs. 9, 20), inclusion type and dis­tr ibution (Refs. 7, 10, 16, 21-26), strain aging (Refs. 22, 23, 27, 28), banding (Refs. 9, 29, 30), hydrogen (Refs. 21 , 23, 28, 31), heat input, and

Poor ST /Duct i l i ty '

/ l /el i.n,. Deoxidation

t ond Rolling " \Practice

y-Type Non-Metall ici^i--Size -Inclusions r ^ - Shape

Spacing •• Banding

Matrix U < ----- Texture Properties ' '^— Grain Sizer/Comp3sil

^ Structure NHeot Tmt r K ..,,. i - *" Strain Aging Embrit t le-L—'^ r. 1« , k~. -Dissolved Oxygen ment I - - _ , , _, ' * Hydrogen

High S T / Strain \

Restraint L==r—— Structural Restraint y " Strains H "*• Joint Contraction

Thermal i __-—- Thermal Gradients Strains 1 " ~ Internal Constraint

N^ Transforma-i,^--Compositicn tion Strains P ^ - Cooling Rate

Fig. 2 — Variables in lamellar tearing

H=45KJ/m{l773KJ/m)

V= 28 5 Volts

1 = 310 Amps

S = I2 ipm (5 08 mm/sec)

;T l d.

(URL) . (do-dl) ( t -do)*

1 distance from the top of the i weld

layer to the top of the cantilever plate

Fig. 3 — Welding sequence

Fig. 4 — Failure close to the plate surface despite the opening up of centerline lam­inations in Si killed A515-65 (Y) steel

preheat (Refs. 19, 22, 27, 31). Uncer­tainty also exists about the mechan­ism of tearing, its location (Refs. 20-22, 32) as well as the t ime and tem­perature of its occurrence (Refs. 17, 20, 23, 27).

According to some investigators (Refs. 20, 27) lamellar tearing is an elevated temperature phenomenon occurr ing in the temperature range 200-300 C while others (Refs. 17, 23) report it to be a form of delayed cracking occurr ing at room temper­ature a few hours after welding. The tears are reported to initiate at the toe or root of the weld or anywhere within the base metal f rom the fusion zone to the center of plate thickness.

Attempts to correlate lamellar tear­ing susceptibil ity to through-thick­ness tensile and notch ductility, in­c lus ion d is t r ibu t ion , or u l t rasonic measurements have not been very successful, partly due to a lack of quantitative data on the tearing sus­ceptibil ity and partly due to com­plexity of the phenomenon itself.

In an at tempt to al leviate this situation a weldabil ity test was de­veloped at Lehigh (Ref. 22) to pro­vide a quantitative measure of sus­ceptibility to lamellar tearing (Fig. 1), The test consisted of joining a can­tilever to a rigid vertical test plate by deposit ing a multipass weld in a 45 deg bevel groove while maintain­ing constant levels of through-thick­ness stress by externally loading the cantilever. The stress level just neces­sary to cause failure during testing, without the need for an overload, designated the critical weld restraint level (CWRL), was chosen as the criterion to represent the lamellar tearing susceptibil ity.

In the present investigation, the mater ia l var iables cont ro l l ing the cracking phenomenon were studied by the Lehigh test to determine the susceptibil it ies of a range of steels and to examine the roles of material p a r a m e t e r s and f r a c t u r e m e c h ­anisms in tearing.

Var iables in Lamel lar Tear ing

In a welded joint, lamellar tearing may be expected to initiate when and whe reve r the t h r o u g h - t h i c k n e s s strain exceeds the ductil ity. While the overall weldment strain depends on the degree of joint contraction and the restraint intensity of the sur­rounding structure, the local stress-strain distributions are highly hetero­geneous and result f rom the com­plex interaction of weld thermal gra­dients and the constraint imposed by the sur round ing con t i nuum. The through-thickness ductility is gov­erned by metallurgical variables such as type, size, shape and spacing of

nonmetall ic inclusions, laminations, propert ies of the intervening matrix, c rys ta l lograph ic texture, band ing , and embritt l ing mechanisms that may act during (or after) welding to lower the local ductility. These, in turn, are influenced by a variety of factors such as melt ing, deoxidation and rolling practice, material composi t ion, heat treatment, presence of hydrogen, dis­solved oxygen, strain aging, struc­tural restraint, welding condit ions, etc. The variables controll ing the tear­ing phenomenon are schematically illustrated in Fig. 2 and, among them, the elongated nonmetall ic inclusions are probably the pr ime cause of poor through-thickness ductil ity.

Nonmeta l l ic Inclusions

The reported (Refs. 10, 15, 22, 26) lack of a direct correlation between the inclusion content and lamellar tearing susceptibil ity suggests that ef­forts to achieve steel cleanness are not necessarily the most economical methods of achieving improvement in lamellar tearing resistance. The cost and cleanness of steel are influenced by the route fol lowed during steel making and the best route is one which would provide, at a min imum cost, the least harmful inclusions as regarcs their type, size, shape, and distr ibut ion.

The shape and size of inclusions after forming operations depend on their formabil i ty characteristic which is in f l jenced by their composit ion and the working condit ion. Silicates are not deformable at room temper­ature but deform at higher temper­ature, the extent of which depends on their compos i t ion (Refs. 33, 34). Spinels, alumina and calcium alu-minates do not deform at temper­atures encountered in steel rolling and re ta in the i r g l o b u l a r shape whereas the monophase type III MnS are highly deformable and elongate to form stringers (Refs. 33, 34).

Depending on the degree of de­oxidation, sulfide inclusions may be either type I, II, or III. While type III sul­fides are formed in steels thoroughly deoxidized with an excess of alu­minum, the type I oxysulfides occur in r immed, semiki l led or sil icon kil led steels where the oxygen content of the liquid steel is relatively high (Refs. 35-37) The type II sulfides occur as a dendrit ic network in steels having an in termedia te oxygen content and form closely connected groups of elongated inclusions after roll ing. In­creasing addit ions of zirconium and rare earth metals lead to a modif ica­tion of the sulfide morphology, the in­clusions formed tending toward equi­axed angular particles or spheres (Ref. 35) . Un l i ke t ype I I I , t hese

342-s I N O V E M B E R 1 9 7 6

'al loyed' sulfides resist deformation and retain their globular shape. Injec­tion of si l icon-calcium, a luminum-cal­cium or misch metal under a highly basic slag is reported (Ref. 38) to reduce sulfur contents to below 0.005 percent.

In Si killed steels, the major inclu­sions are the mult iphase silicates, while the minor ones are type I sul­fides and simple oxides. With simul­taneous Si-AI deoxidation the inclu­sions formed display a wide range of composit ion and morphology due to complexity of the reactions and are of the type M x A y S z (manganese alumi­num silicates), A y S z (mullite) and a mixture of type I and III sulfides (Ref. 39). If Al deoxidation is carried out before Si addit ion, the major inclu­sions are the monophase type III sul­fides and the minor ones are alu­m ina , sp ine ls and c a l c i u m a lu -minates. One can, thus, anticipate lamellar tearing to initiate at the sil­icates in Si killed and semikil led steel and at the sulfides in an Al killed steel.

The propensity for void formation depends on the differential thermal expansion characteristics between the inclusion and steel matrix. Where the inclusions contract more than the matrix, as with sulfides and iron ox­ides, voids tend to form; but if the in­clusions contract less than the matrix, as with silicates, tessellated stresses are generated (Refs. 40, 41).

Among the techniques available for inclusion count ing, the Quantimet is reported (Refs. 42, 43) to give accu­rate and reproduc ib le measure­ments. Nevertheless, at the mag­nifications normally employed for the assessment of steel cleanness, it is not possib le to account for the smaller inclusions — especially the oxides — which may exert significant influence on the lamellar tearing sus­ceptibility. In some steels, especially the Al killed type, longitudinal streak­like defects containing aligned clus­ters of small alumina inclusions occur which apparently result f rom reox-idation during teeming and get en­trapped in the ingot during solidif ica­tion. Such defects may not be in­cluded in the Quantimet data and the steel quality may be overestimated.

Beside the effects of inclusions and laminations, other factors like band­ing, c rys ta l lograph ic tex ture and matrix embrit t lement may play a sig­n i f i c a n t r o l e in e n h a n c i n g t h e mechanical anisotropy of rolled prod­uc ts . Wh i l e the s i gn i f i cance of banding with regard to lamellar tear­ing is not es tab l i shed yet, the through-thickness ductil ity is report­ed (Refs. 29, 45) to drop with an in­crease in the degree of banding. A homogenizing treatment has been

reported (Refs. 15, 16, 26, 29, 45) to eliminate the effects of banding, but Jatczak et al (Ref. 46) and Grange (Ref. 47) conc luded that homo-g e n i z a t i o n c a u s e s on l y s l i g h t , commercial ly insignificant improve­ments in the transverse ductil ity and impact strength.

Some steels, especially the con­trolled rolled type, have been re­ported (Refs. 48-50) to exhibit peri­odic splitting parallel to the rolling plane and this has been attributed to cleavage along textured ferrite bands (Ref. 48), tearing along deformed fer­rite grain boundar ies conta in ing numerous carbides (Ref. 49) and sep­aration along embrit t led sulfide inter­faces (Ref. 50). Splitt ing is reported (Ref. 48) even among the rare earth treated steels containing low sulfur levels with shape control. Also, the habit planes of mar tens i te laths, formed from unrecrystall ized aus­tenite, tend to have preferred orienta­tion due to the texture in the parent austenite and this may contribute to splitting in the matrix.

The matrix l igament separating the inclusion voids may be embri t t led, either during or after welding, by fac­tors such as strain aging, dissolved oxygen, hydrogen, etc., thus facil­itating lamellar tear initiation. Jubb et al (Ref. 27) have shown that in mult i­pass welds, the HAZ of the parent material passes through a temper­ature range where limited ductility combined with a tendency to strain aging is likely to be experienced. The maximum strain aging is found to de­velop at temperatures in the range of 250-300 C. The strain aging may rapidly embr i t t le the matr ix l iga­ments separating the inclusion voids so that these l igaments fracture at low strains. As little as 20-30 ppm of ox­ygen in solution is reported (Ref. 51) to produce a marked intergranular embrit t lement of iron without a sig­nificant rise in the yield stress and in­creasing quantities of oxygen are f o u n d to be p rog ress i ve l y e m ­britt l ing.

There is limited evidence (Refs. 21 , 23, 28) to suggest that lamellar tear-

A l - k i l l e d Mil 16113C(T)-1" S i A-285(M)-• i • ••'..

\

Root

S i - k i l l e d A515-70(G)-2 1/2" A l - k i l l e d Mi ] U 1i 3C(J) -2" S i - k i lied A-36(Y)-2"

Fig. 5 — Location of lamellar tears in various steels. (Reduced 33%)

WELDING RESEARCH S U P P L E M E N T ! 343-s

CO

CD C

5

<D

o i q 1 o

a CJ

o

(1) - CD

t 0 2 r -O CO T -

o o o 1 o d d

0) _ . _ - < » O > r- O § CO UO CO CO 2 O r - O O CO O O O O O

O O O O O

> r - >

I I I I I

O

- - O o CO IT) N- t— T - - 5 s t n t M i O t - ^ O O O O O CO

d o d o d o

O

o O ) O O r -O O

t n CM co CO "0" -M" CM

O v- o o o o o o

C M S r -CV C O N

o o o o o o

T - CO 0 0 CO

o o o o

CO t - CJ) CO CO CM CO OJ

o o o o o o o o

o LO LO

o o o o

o o r~ •* CO N ^ LD

o o o o o o o o

CM O ) CM C O I O N

o o o o o o

CO I -r - • *

o o o o

C V O T T -

S LO CD CD

o o o o o o o o

r*- CM o o o o

LO CO CM S

O O O r-O O O o

CO CM LO

i - o o o o o

CM CM T - O O O

' t N W

o o o -a-o o o o

CO t 1 CJ

V V

LO LO O O O O

LO LO LO LO O O O O

O O o o

LO LO CM O O O O O O

LO O O 1 -

O o

CO t S CN LO t f ^ f CO O o O O

V V V V V V V V

o 5 r-J O

LO CM

oo r~ LO ^r i - O O CNJ

•M-CM i - CO

CO o o

CO T - O CM O

CO LO LO 0 0 CO t - > - O C V

o

(/) o Q E 0 u

o

CO CM

o o

CM - i — T - O

N CO CM i n O O O T -

LO N I N CO i - © O t -

[N . * 1 -

o o o V

CO T - CM

i - o o

O ) T -

o o V

* t f T -i - o

C O O N C O O r - 1 - O

i - i - o n T - T - CM O

r- r— o> •* i -CM CM r~ CM CM co o o i - o o o

i - L O O CM 1 - r-O O O

CM T -O T -O O

LO LO LO CO

O O o o o o o o

CM T - CO O 7 - T -O O O

o

>, n

X E

o

ttT

a

^-B

cl> T5

0

o

W CM CM

o>

ra " ' •

LL

s

CO o CM =

' OT

-a CD

o a <D DC

X

^T CM

* o c

JxC

c_ J

o CJ

CD

ra

LL.

z o

i n CM

a cl) X4

0

'.) CU LL

I I

CO CM

• a CD

c o Q . CD

rr

>

O) C\l

11 CD

•f-0 a. CD

LL

V

r— "» CM r-

CD

X3 3 CD r -

T_ ca o * ; a j a CD CO

r r LL

> - CL

<ji i - i - co cn o T - CM CNJ CM i - CO o o o o o o

•si- 1 - o> co - i - o O i - O O i - i -o o o o o o

CD

CD

W M- CO CO CO - Q CM CO h - LO

CD r -

o t r r-N O C O

1 - T -

CD

B M - M " W CM O ) ^ t r LO LO - Q CM O CM CO

a i r - r - r - r -

O CT) 7=- -sj- CO LO LO C\J CNJ •— CSJ CNJ i - i -

• • £ ' • • •

ra 1

cn LL

ra 11 Cl 01 L L

• D C CD g E O O c a . . * CD C

t r 3

-o TO 0 0 T— T-O O Q . Q . CD CD

CC cc

a 11 a m

LL.

a ai

a

CD

CO

1 0)

JO CO

h-CD

C c

"ra

a> ra

ci

^ ^

o LO CO CO CM CO < <

r - CM

CO CM T - CO

CD —

JO o o io

• i n L O i r>

LO LO LO

u_ < < <

CO -st LO CO

CNJ CNJ £

5 CO LO CO T - CO i

to co X < < LU

h - 0 0 a>

CO LO

a a CO CO CO CO CM CM < <

O i -

CO LO r - C M r - r - T -

O O CO CO

O r- T - T -r^ r - i - T -

* i - ^ CD CD co CM ir LO LO = =

< < 5 5

CM CO td - LO

a C3 0 0

LO

<

CD 00 Z- 00 CO I— CO

CO P 111 a)

CO

rr

CT 0 0 CO LO

<

<1> CD

to

N

CD O

E 3 t a a>

e to S S E S

i i ' Z ° cc B a c 3 2 1

t-

jh

<I>

E

7*J

4)

in

3 4 4 - S | N O V E M B E R 1 9 7 6

ing is a form of delayed cracking oc­curring at room temperature a few hours after welding. Several in­vestigators (Refs. 52-55) have report­ed on the beneficial effect of sulfides which, acting as hydrogen sinks, tend to lower the hydrogen activity in the lattice and, thus, reduce the likeli­hood of underbead cracking. How­ever, if voids and microcracks are formed around inclusions, as is the case in lamellar tearing, the presence of occluded hydrogen may increase the likelihood of subsequent prop­agation by hydrogen assisted crack­ing (Refs. 55, 56).

Materials and Procedure

A range of materials which includ­ed two semikilled, eight silicon killed, six aluminum killed, one rare earth treated and one leaded-sulfurized steel, with the thickness ranging from 18 mm to 400 mm (3/4 in. to 8 in.) were tested to evaluate their tearing susceptibility. These steels were sup­plied by various manufacturers and fabricators — some coming from por­tions of a structure that had dis­played lamellar tearing during actual fabrication, some belonging to grades that have had reported his­tories of lamellar tearing and some that have had no known history of lamellar tearing. The material details are presented in Table 1.

The lamellar tearing susceptibil­ities of the various steels were deter­mined with the Lehigh test pro­cedure outlined by Oates and Stout (Ref. 22). This test was developed to provide a simple and effective means of evaluating the lamellar tearing sus­ceptibil ity of various materials through the application of an exter­nal load to supplement the weld ther­mal and restraint stresses. The exter­nal load was increased proportion­ally to the cross sectional area of the weld after each weld layer had cooled to room temperature so as to main­tain a given level of through-thick­ness stress, called the weld restraint level, which was characteristic of each test. The load required was cal­culated using the simple formula given in Fig. 3 and the restraint level just enough to cause failure during welding, called the CWRL, was used as the susceptibility criterion.

In determining this criterion, care was taken to see that plates less than 38 mm thick were stiffened by weld­ing a 25 mm thick back-up plate to them so that the longitudinal stresses due to bending between the restrain­ing bolts was small compared to the through-thickness stresses. Thus, the CWRL could be related to the mate­rial parameters without having to con­sider the variable effects introduced

through changes in the plate stiff­ness (thickness).

An automatic gas metal-arc (GMA) welding process was used and the welding conditions employed are listed in Table 2. The extent and nature of subcritical tears were studied by metallographic examina­tion of sections taken from the spec­imens tested just below the CWRL and microhardness measurements were carried out at regions around the tears to gain additional under­standing of the factors influencing tear location.

The inclusion distributions in the various steels were evaluated using the Quantimet 360. Areas approx­imately 5 mm below the plate sur­face were scanned as the tears have been observed to lie mostly in this zone. Fifty fields were scanned in each section and a minimum of four sections were examined for each material so as to give adequate sam­pling. The average values of area fraction (A), aspect ratio (A ) , num­ber of inclusions (N), inclusion length (L) and the longest inclusion in each field (L x ) were determined. The inclusions were identified both from the polished sections, using the elec­tron microprobe, as well as from the fracture surface, using the energy dispersive spectrometer (EDS) asso­ciated with the SEM. The oxygen con­tents of the steels were determined by the vacuum fusion method and repre­sent both dissolved and combined oxygen. Fractographic studies were carried out on the SEM by examining various regions on the fracture sur­face and stereopictures were taken to assist in better interpretation of the fracture features.

Results

Behavior of the Various Steels

The Lehigh lamellar tearing test results are presented in Table 3 and

the susceptibility of the various steels are summarized in Table 4 together with information on their compo­sition, matrix structure, inclusion type and distribution, tear location, frac­ture mode, etc. From Table 4 it can be seen that different heats of steel of the same grade display a wide vari­ation in tearing susceptibility.

For instance, the susceptibility (CWRL) of Si killed A515-70 grade varied from 44 ksi (303 MPa) for the G-heat to 61 ksi (421 MPa) for the F-heat to 71 ksi (490 MPa) for the K-heat. The 80 mm (3.15 in.) thick A515-65 (V) plate showed a high resis­tance to tearing despite a heavy centerline lamination and, though the

I n i t i a t i o n by L iqua t ion in the Fusion Zone of Pb-Sul fur i2ed F.M. S t ee l (N).

I n i t i a t i o n by S p l i t t i n g in c o a r s e Grained HAZ

100X Combination of Lamellar Tear and I n t e r -Granular Cracking in A l - k i l l e d A-533 Gr. B (D)

Fig. 6 — Crack initiation by mechanisms other than lamellar tear. (Reduced 54%)

Table 2 — Welding Conditions

1. Filler metal(a)

2. 3. 4. 5. 6. 7. 8.

9. 10.

Shielding gas Gas flow rate Current Voltage Travel speed Electrode extension Inter-pass temperature Number of passes Heat input

Aircomatic A-632, 0.045 in. (1.143 mm) UTS = 123 ksi (848 MPa) YS = 109 ksi (752 MPa) % Elong = 22 RA = 65 CVN at - 50 C = 70 ft-lb (95 J) Argon + 2/02

50 cfh (23.6 l/min) 315A 28 V 12 ipm (5 mm/s) 3A in. (19 mm)

Room temperature Nine (9) 45kJ/in. (178kJ/m)

(a) Composition: .07 C, 1.35 Mn, .50 Si, .012 S, .01 P. .45 Mo, 1.30 Ni, .15 V

WELDING RESEARCH S U P P L E M E N T ! 345-s

Table 3 — Results of Lamellar

Material

F.M. Grade (N) 1 % in. f rom 3% in. plate

EH33 (P ) % in.

(a) A516-70 (E)

1 in.

A285-C(M) VA in.

Mi l -24113C(U) 1 in.

A515-70(F) 1 in.

A516-70(R) 2% in.

Mi l -16113C(T) 1 in.

A-36(Y) 2 in.

A-36(X) 114 in.

A515-65(V) 3VA in. laminated

A533Gr .B (D) 2 in. f rom 5% in. pi.

A515-70(K) 21/2 in.

A 5 8 8 G r . A ( W ) 1 in.

Z Steel (S) ~\V2 i n .

A283Gr .D 3 in. f rom 8 in. plate

A283 2% in. f rom 5 in. plate

Speci ­men

7NA1 7NA2 7NA3 7NA4 7NA5 7PA1 7PA2 7PA3 7PA4 7EA1 7EA2 7EA3 7EA4 7MA1 7MA2 7MA3 7MA4 7UA1 7UA2 7UA3 7UA4 7FA1 7FA2 7FA3 7FA4 7RA1 7RA2 7RA3 7RA4 7RA5 7TA1 7TA2 7TA3

7TA4 7YA1 7YA2 7YA3 7YA4 7XA1 7XA2 7XA3 7XA4 7VA1 7VA2 7VA3 7DA1 7DA2

7DA3 7DA4 7DA5 7KA1 7KA2 7KA3 7WA1 7WA2

7WA3

7SA1

7BA1

7BA2 "CA1

Teari

ksi

55 40 28 36 35 50 44 46 42 55 50 45 45 52 60 56 48 64 60 57 56 65 50 60 58 60 72 64 62 58 56 62 68

64 60 68 64 66 56 68 72 68 64 70 67 56 52

76 72 65 52 72 68 56 76

72

60

52

72 60

ing Tests

WRL (MPa)

(379) (276) (193) (248) (241) (345) (303)

' (317) (290) (379) (345) (310) (310) (359) (414) (386) (331) (441) (414) (393) (386) (448) (345) (414) (400) (414) (496) (441) (428) (400) (386) (428) (469)

(441) (414) (469) (441) (455) (386) (469) (496) (469) (441) (483) (462) (386) (359)

(524) (496) (448) (359) (496) (469) (386) (524)

(496)

(414)

(359)

(496) (414)

Fail­ure

Yes Yes No No No Yes No Yes No Yes Yes No No No Yes Yes No Yes Yes No No

Yes No No No No Yes Yes Yes No No No

Yes

No No Yes No No No No Yes No No Yes No No No

Yes Yes No No

Yes No No Yes

No

No

No

No No

Fracture stress

ksi

— 37.5 37

— — 46 45

— — — 51

— 53

— — — — — 58

— — 62 61 53 64

— — — — 67 67

—

— 67

— 67

— 68 69

— — 70

— — —

no 1 at 74

— — 70 72

— 71 77

—

75

77

83

>75

>75

(MPa)

— —

(259) (255)

— —

(317) (310)

— — —

(352)

— (365)

— — — — —

(400)

— —

(428) (421) (-365) (441)

— — — —

(462) (462)

—

— (462)

— (462)

— (469) (476)

— —

(483)

— — —

ailure (495)

— —

(483) (496)

— (490) (531)

(517)

(531)

(572)

(517) (517)

Remarks

Failed, loading after 3rd pass Failed, loading after 6th pass

— —

Metal lographical ly examined Failed, loading after 6th pass

— Failed, loading after 9th pass Metal lographical ly examined Failed, loading after 8th pass Failed, loading after 9th pass

— Metal lographical ly examined

— Failure after 6th pass Failure after 6th pass Metal lographical ly examined Failed, 10 h after loading Failed, loading after 6th pass

— Metal lographical y examined Failed, loading after 7th pass

— —

Metal lographical y examined

— Failed after 2nd pass Failed after 4th pass Failed, 10 h after loading Metal lographical ly examined

— —

Failed, loading after 9th pass at 64 ksi

Metal lographical ly examined

— Failed after 4th pass

— Metal lographical ly examined

— —

Failed, loading after 8th pass Metal lographical ly examined

— Failed, loading after 9th pass Laminations opened at center Failed in weld after 7th pass Metal lographical ly examined

Failed along HAZ after 3rd pass Failed along HAZ after 5th pass

— —

Failed after 24 h a t 72 ksi Failure after 6th pass

— Failed while restor ing load to

76 ksi, 15 h after welding Failed while over loading 18 h

after welding Failed in the weld metal

Failed near HAZ

Metal lographical ly examined No fai lure but ultrasonic

indications after overloading

CWRL ksi

37

(a) 5 0 "

53

58

61

62

66

67

69

69

70

71

75

> 7 5

> 7 5

>75

(MPa)

(255)

(345)

(365)

(400)

(421)

(428)

(455)

(462)

(476)

(476)

(483)

(490)

(517)

(517)

(517)

(517)

(a) Tested without stiffener

346-s I N O V E M B E R 1 9 7 6

Fig. 7 — Array of parallel cracks in semi-killed A-285 (M) steel

r~W MS. '^^^k J^-

T * S —

mViiJ.iiija.jp urn

Fig. 8 — Terrace crack through a ferrite band in Al killed Mil 24113C (U) steel

*kiiJmJ&i& *****

Fig. 9 — Fracture surface showing lamel­lar tear initiation (right) and cleavage propagation (left) in semikilled A-285 (M)

laminations opened up under load (Fig. 4), fai lure occurred close to the plate surface, suggesting that tear­ing suscept ib i l i ty is most ly in f lu­enced by the material condit ion close to the plate surface.

A m o n g the mater ia ls that had exhibited lamellar tearing dur ing ac­tual fabricat ion, the ABS EH-33 (P), A-285 (M) and A515-70 (G) d is­played a high suscept ib i l i ty ( low CWRL) in the Lehigh test whereas the A-283 (heat B and C) steels showed a low susceptibil i ty (high CWRL). This showed that a material with inher­ently high resistance to tearing may suffer lamellar tearing, but only if the res t ra in t c o n d i t i o n s a r e seve re enough. The rare earth-treated Z-steel exhibited a high resistance to tearing and failed in the weld metal rather than the base metal, thereby demonstrat ing the beneficial effects of rare earth addit ions on the tearing susceptibil ity. There was no definite trend in the susceptibil i ty with vari­ations in the plate thickness.

Nature and Location of Tears

In the Lehigh test, the tearing tend­ed to initiate while the arc was mid­way through the pass, in the end por­tion of the plate, adjacent to the toe of the preceding layer. This was due, apparently, to a transient overload occurr ing at the end port ion of the weld from a reduction in the net sec­tion support ing the external load as the molten puddle entered the junc­tion of start tab and test plate. Metal­lographic examinat ion of subcrit ical tears showed lamellar tears to ini­tiate anywhere f rom the lower part of the HAZ to locations up to 13 mm below the plate surface (Fig. 5).

However, in EH-33 and the free machining steel the crack initiated close to the fusion zone, apparently by l iquation, and propagated as a lamellar tear (Fig. 6). In A533 Gr. B steel, tearing occurred in the coarse grained HAZ by a combinat ion of intergranular cracking and lamellar tearing (Fig. 6).

For the same plate thickness, the

depth location of tearing tended to be maximum in the Al kil led steels (Mil 24113C, Mil 16113C), intermediate for the Si kil led steel (A515-70) and m in imum in the semik i l led steel (A285) (Fig. 5). In the semiki l led and Si killed steels, extensive tearing was observed outside the main crack front (Fig. 7) but not so in the Al killed steels. In banded steels there was a tendency for the terrace crack to run within the ferrite bands (Fig. 8).

The fracture surface (Fig. 9), often displayed a region of ducti le lamellar tear initiation and a region of brittle cleavage propagat ion, the propor­tion of which varied depending on the material parameters and the test con-d i t ion . In Al k i l led A516-70, Mi l 24113C and Mil 16113C steels, 100% ducti le fracture was observed (Fig. 10), while other steels displayed vary­ing proport ions of lamellar tear and cleavage.

On a microscopic scale, lamellar tears were characterized by a step­like pattern with stacks of terraces linked together by shear walls (Fig. 11). The terrace cracks displayed an array of elongated voids whose size and density varied among the differ­ent steels depending on their sus­ceptibility (Fig. 12). The voids tended to be larger and denser in the more s u s c e p t i b l e s tee ls . T h o u g h the terrace was generally associated with the inclusion voids, in some steels they showed the following features (Fig. 13): (a) splitting along lamina­tions, (b) network of a l iquated phase along prior austenite grain bound­aries (only in EH-33 and F.M. Steel), and (c) clusters of equiaxed dimples containing oxides (MnO, alumina, etc.) and sulfides (MnFeS, Type I MnS).

In EH-33 (P) steel, the terrace dis­played a dense network of a l iquated phase which was enriched in Si, Al , Mn, Fe, S, Nb, Ca and Ti (Fig. 14) and in the leaded-sulfurized steel (N), the terrace was characterized by both elongated sulf ide inclusions and a dense network of FeS containing clusters of small (MnFe)S inclusions (Fig. 15).

Fig. 10 — Fracture surface showing 100% ductile lamellar tear in Mil 24113C (U)

Fig. 11 — Stacks of terraces linked together by shear walls in Si killed A515-70(F)

Inclusion Characteristics

Typical inclusions within voids on the fracture surface of some steels are shown in Figs. 16, 17 and 18, and inclusion distr ibution in various steels are shown in Fig. 19. While the inclu­sion types present were chiefly influ­enced by the deoxidation practice, their composit ion varied with the rela­tive proport ion of deoxidizing ele­ments present.

For instance, in the semiki l led A-285, which had a low Mn content (0.41), the oxides and silicates were found to be highly enriched in Fe, whereas in the semikil led A-36, con­taining nearly twice the Mn (0.78) and some Ce (0.01), the proport ion of Fe in the oxides was much less. Silicate

WELDING RESEARCH S U P P L E M E N T I 347-s

CD Q -

-o o

I-o

8° 11 co

'c

1

c o

ro 3

cr

_ i

c p> o £ CO — 3 CL

CTOT

CD

Dl 03

> ra OJ

CD

Ol co > CO CD

CD

o> ra > ra CD

CD CT ra > ra CD

CD CT ra > (0 CD

o o

CD

ra

o o o o

ra O

CJ

C) > :>

ra > ra 0

n > >

o

o

RE

35

.O.

o

1

i i XI ra

x ^

C -a CD C

O > > o >

o

ro u o

_ l

U)

"o ra CD

N < X _ cp

a a Z>

N < I ^ CD

Q. Q .

Z>

o CO

.Ew

i ® 5

CD • CD

a.co a >, . •> ,

I - 0 . ( -

— CD

rr. "° a — • > , - >

h- CO

CO

ro

L

CO

Q -

+ CO

-a o>

I " £T3

< X

CD

g o

< I

CD g o

N < X g _o CD

CQ

N < X g C^

CD CQ

N < X

CD g O

N < X g CD

CQ

N N c < <

-ill CD CD

I - Q CD

N

• * g CD

^ 0 a CD C L

< X

g o

CD

I SS

T - ^ r c n

co c o T - : CO CM CM a> CM T ^

o •

Q . O Q . ^ > , CD > ^

2 ° = CD

r~. r-

CD T3

Q . 2

T3

B CD

L : G CD

CO C D

o .>• II

cn E CD II

a. +

CO u _

CL

+ C L

- o co§ S o I, c3 « -o ™

>. II * " li

o •a a.

-a n ID >•.. ^ II S £ c o 2 c o .£co E ° ECD ^CD g, II g, II g> II ° a: ° n - a CL

1 o c CD c CD CO > CQ LU

o LU cr UL cr

CT

ro ro CD a r~ •Q co 8 co S: II Lt 0

CT C

C

ra

CL

+ C L

+

CD

T3 O) Cp LO

S) II § °- -g Q.' I Li. o U 5 c D n O c + c , t + > C D > C D o C D £ C D = > C ! O u O u O u o T | j + „ CD CD O 0 (j-

LL OT OT 2 I - HI

"D 0 CO

I I CO l : CO II

o-° LU

W f .

2 « LL

O0 CO

T LU

CD LL — o o

co SS co ^ LO ^ LO : CM T - CO

< < CM ;

UJ L.

< <

> X. CJ — — CO CO T- ; -" . 1 -C O . S . > T-

L O ^ L O ^ c O a C M : ' T C O ' T r C M C O C M — r

11 ,T — O r - CO N- . i -

<° 5? CO = - i - CM — T

cn

O . ^ CO ^ 00 .

- CO LO °0 T~

0 = £ S

348-s I N O V E M B E R 1 9 7 6

* J> >"W

Fig. 12 — Array of inclusion voids on the terrace of various steels, (a) Mil 24113C (U): (b) A516-70 (E); (c) Mil 16113C (T); (d) A588 (W). (X200, reduced 52%)

Fig. 13 — Types ot terrace cracks: (a) splitting along laminations: (b) network of a liquated phase: (c) cluster of equiaxed dimples containing alumina: (d) array of elongated voids. (Reduced 52%)

Fig. 14 — Liquated phase in EH-33 (P) steel Fig. 15 — Liquated FeS network containing (Mn, Fe) S inclu­sions, in Pb-sulfurlzed F. M. steel (N). (Reduced 41%)

inclusions were mostly in the form of stringers (Fig. 19a) whereas the sul ­fides were present in a wide range of shapes and sizes. Type I MnS was of­ten d u p l e x wi th a eu tec t i c ta i l surrounding the monophase MnS and Type III MnS was generally pres­ent as a fine stringer (Fig. 19b). Clusters of small aluminate inclu­sions with polyhedral contours were found in the Al killed steels and, in some cases, they appeared as a con­tinuous trail containing hundreds of inclusions (Fig. 19d).

The inclusion count obtained with the Quantimet did not show a satis­factory correlation with the lamellar tearing susceptibility, as indicated by the CWRL. Though suscept ib i l i ty tended to be low among steels with a very low inclusion count — A515-70

(K), A588 (W), A533 Gr. B (D), Z steel (S) — some steels such as, A36-(X), A36-(Y) and Mil 16113C (T), dis­played a low susceptibil ity despite their high inclusion count (Table 4).

Resul ts f r o m oxygen ana lys is showed lamellar tearing tendency to increase with an increase in the ox­ygen content of the steel (Fig. 20). However, the semikil led and Si killed A-36 steels displayed a high resis­tance to tearing despite their high ox­ygen content.

Mechanism of Tearing

The progressive stages in lamellar tearing are shown in Fig. 21 for the Si killed A515-70 material. The first stage was void formation which oc­curred either by decoherence of the

i n c l u s i o n - m a t r i x i n te r face or by cracking of inclusions. The second stage was the linking of adjacent voids on the same plane by the rup­ture of intervening matrix to form a terrace crack. And the final stage was the linking of terrace tears on adja­cent planes by ver t ical shear ing which gave the characteristic step­like appearance to the fracture.

While the smaller inclusions tend­ed to separate by interface deco­herence (Fig. 22) those with large aspect ratio often displayed multiple internal fracturing (Fig. 23). Silicates invariably shattered within the voids whereas the sulfides tended to de­cohere, unless the aspect ratio was very high (Fig. 24). Larger inclu­sions tended to form voids earlier than the smaller ones (Fig. 25). The

WELDING RESEARCH S U P P L E M E N T , 349-s

extent of void growth around inclu­sions var ied among the di f ferent steels and tended to be greater in materials exhibit ing a high resis­tance to tearing (Fig. 26).

When the inclusions were closely spaced, the voids tended to coalesce by necking (Figs. 27a, 28a) whereas when the inc lus ions were farther apart , the vo ids general ly l inked together by a sheet of microvoids containing small oxides and sulfides (Figs. 27b, 28b). The proport ion of microvoids varied among the differ­ent steels and tended to be higher among the more susceptible steels. For instance, among the semikil led steels, A-285, which was more sus­ceptible than A-36, contained a far greater proport ion of microvoids on the terrace than the A-36 (Fig. 26). Similarly, among the Al kil led steels, the proport ion of alumina clusters was found to be considerably higher in Mil 24113C (U) than in Mil 16113C (T) (Fig. 13c). The latter steel dis­played a greater resistance to tear­ing despite its higher inclusion count (i.e., type III MnS).

In a few rare instances, void link­age occurred in a "z ig-zag" manner with ridges and valleys (Fig. 22c) and,

in one material (A-533 Gr. B), the voids were joined by intergranular separation (Fig. 27d). Void linkage also occurred by quasi-cleavage and cleavage (Fig. 28d). In materials ex­hibiting cleavage propagation, islands of inclusion voids were often seen surrounded by cleavage facets, sug­gesting that voids had formed earlier in the fracture process while the lig­aments connecting them cleaved dur­ing the final unstable propagation (Fig. 29). The vertical shear linkage of terrace cracks occurred late in the tearing process and was associated with microvoids (Fig. 30).

Discussion

Investigation of lamellar tearing susceptibil ity on a wide range of materials has shown that susceptibi l­ity to tearing is a function of many variables and cannot be generalized on the basis of steel grade, plate thickness, deoxidation practice, etc. This was evident f rom the wide vari­ation in the susceptibil ity observed among different heats of steel of the same grade and also from the lack of

Fig. 16 — A large silicate surrounded by microvoids containing Fe rich oxides and oxysul-fides in semikilled A-285 (M) steel. (X2000. reduced 42%)

corre la t ion between suscept ib i l i ty and plate thickness, inclusion type, or inclusion count.

This behavior was understandable from the fact that lamellar tearing is a complex phenomenon governed by a variety of mechanical and metal­lurgical factors.

Meta l lograph ic examinat ion of subcritical tears in the various steels clearly showed lamellar tearing to oc­cur in three distinct stages, namely: (a) void initiation, (b) void growth and (c) void l inkage.

Void Initiation

This was observed to occur either by the decoherence of inclusion-matrix interface or by the cracking of inclusions. The latter effect was con­fined to large silicates or sulfides (Figs. 23 and 24). This may be be­cause, as indicated by other inves­tigators, (Refs. 57-61) larger inclu­sions tend to contain more flaws and, hence, are susceptible to internal fracture while the smaller inclusions, remain intact and nucleate voids by inter face decoherence (Fig. 22). Argon et al (Ref. 62) have shown that for decoherence to occur, the local stress must exceed the interfacial strength. The magnitude of the inter­facial stress is reported (Ref. 62) to be influenced by factors such as resid­ual t esse l l a ted s t ress , i nc lus ion aspect ratio, matrix flow strength and inclusion interactions, while the inter­facial strength is dependent on the type of Inclusion and the presence of impurit ies such as Sn, P, Sb, As, etc. which tend to embrit t le the interface (Ref. 63).

The larger inclusions were ob­served to form voids earlier than the smaller ones as was evident from the presence of undamaged smaller in­clusions lying beside a fractured or decohered larger inclusion (Figs. 23, 24). This fact implies that, at a given volume fraction of inclusions, uniform

a^^inV' =3f"

Fig. 17 — A large type III MnS beside a small type I oxysulfide in Al killed Mil 24113C (U). (X2150, reduced 40%)

Fig. 18 — Voids containing alumina in Al killed Mil 24113C (U). (X1840, reduced 40%)

350-s I N O V E M B E R 1 9 7 6

distr ibution of smaller inclusions is better than random distr ibution of larger inclusions.

While, in most cases, lamellar tear­ing initiated by void formation around inclusions, in some instances tearing was observed to occur by low energy mechanisms such as splitt ing along ferrite bands, l iquation in reheated HAZ, and in tergranular c rack ing . Thus, even a clean steel may readily init iate tear ing accord ing to the condit ion of the matrix and its sus­ceptibility to HAZ embritt lement.

Void Growth

The extent of voids around inclu­sions was observed to vary among the various steels and tended to be greater among those displaying a high resistance to tearing (Fig. 25). Void extent is a measure of plastic work done and is influenced by the inter-inclusion spacing, matrix flow stress, nature of second-phase part i­cles and the stress state. With an in­crease in the level of triaxiality and matrix flow strength, the plastic flow is inhibited in the matrix and tends to concentrate at the voids where the stresses are higher and the con­straint is low. If the inclusions are closely spaced or if the intervening matrix is free of incoherent particles, then the voids continue to grow until they eventual ly l ink by neck ing. However, in the presence of small sul f ides, ox ides or other second phase particles in the matrix, void growth is l imited and linkage occurs prematurely by a sheet of micro-voids.

Void Linkage

The linking of elongated inclusion voids on the terrace was observed to occur by several modes such as neck ing , mic rovo id coalescence, " z i g - z a g " t ea r i ng , i n te rg ranu la r cracking, quasi-cleavage and cleav­age (Figs. 26, 27). The amount of energy absorbed during tearing was dependent on which of these modes was dominant. Necking and micro-void coalescence were the commonly observed modes. In a few rare in­stances linkage occurred by a "z ig­zag" mode of tearing. Such a mode has also been reported by other investigators (Refs. 64-66). Beachem (Ref. 64) has attributed it to a blend of shearing and tearing on alternating shear planes, causing the crack to fo l low a " z i g - z a g " cou rse . Th is mechanism was not observed in the strongly banded plates.

In the quenched and tempered A533 Gr. B (D) steel, void linkage oc­curred by a combinat ion of necking, microvoid coalesence and intergran­

ular c rack ing . The in tergranular mode of l inkage was only observed in the HAZ, possibly due to some form of embrit t lement. The quasi-cleav­age mode of linkage was also con­fined to HAZ and the small cleavage­like facets suggest tearing through martensite laths (Fig. 27c).

The lack of a satisfactory correla­tion between the inclusion count and the tearing susceptibil ity, even among steels containing similar inclusion types, suggests that the smaller inclu­

sions (oxides, sul f ides, etc.), not represented in the Quantimet data, a n d o t h e r f a c t o r s l i k e c r y s t a l ­lographic texture, banding, matrix embritt lement, etc. may exert signif­icant influence on the tearing sus­ceptibility. Thus, the greater suscep­tibility of semikil led A-285 (M), de­spite its lower inclusion count, as compared to semikil led A-36 (X), may be due to the presence of large amounts of smaller oxides and oxy-sulfides which limited the void growth

Fig. 19 — Inclusion distribution in some steels: (a) semikilled A-285 (M); (b) Al killed Mil 16113C (T): (c) Pb-sulfurized F. M. steel (N); (d) Ai killed Mil 24113C (U). (Reduced 47%)

-CWRL, KsKMPa)

Fig. 20 — Oxygen vs susceptibility (CWRL)

WELDING RESEARCH S U P P L E M E N T | 351-S

TYPE 1 MnS 1.56um

Fig. 22 — Decohered type I sulfide

19/jm

Fig. 21 — Progressive stages in lamellar tearing, (a) void forma­tion: (b) terrace linkage: (c) shear linkage. A515-70 (F). (Reduced 22%)

Fig. 23 — Multiple internal fracture in a long sulfide. Note un­damaged smaller sulfides

and caused premature l inkage of in­clusion voids whereas the A-36 (X) steel had relatively fewer oxides and oxysulfides, as is evident from the fractographs shown in Figs. 25a, b.

Thus, suscept ib i l i ty to lamel lar tearing is a function of the energy ab­sorbed during each stage of tearing which, in turn, is control led by a variety of meta l lurg ica l var iab les. While increasing the steel clean­liness is an obvious method of en­hanc ing the tea r ing res i s tance , improvement may also be sought by:

1. Reducing the inclusion size and altering the shape by addit ions of innoculants, stirring of ladle, etc. so as to enhance nucleation.

2. Avoiding silicates and iron-rich oxides by adopting Al killed practice and adding sufficient amounts of Al to preserve an adequate cover until tap­ping.

3. Controll ing the shape and size of sulfides through rare earth addi­tions.

4. Proper choice of rolling condi­tions to reduce mechanical f ibering, c r y s t a l l o g r a p h i c t e x t u r i n g a n d banding.

5. Enhancing the matrix tough­ness of steel by reducing grain size, alloying, etc.

There are, thus, several metal­lurgical options available to produce steels with a high resistance to lamellar tearing and the choice of method is dictated by considerations of cost and criticality of application.

Fig. 24 — Decohered sulfides and a fractured silicate in ABS EH-33 (P) steel tested below CWRL. (X1600, reduced 22%)

Conclusions

1. Lamellar tearing is a complex phenomenon governed by a variety of metallurgical and mechanical vari­ables and resistance to it cannot, therefore, be simply related to the steel grade, plate thickness, or the in­clusion type.

2. Tearing susceptibil ity is mainly influenced by the material condit ion c lose to the p la te su r f ace anc th rough- th ickness proper t ies ob -

19( jm

Fig. 25 — A fractured silicate beside un­damaged smaller silicates in Si killed A-283 (B). (X533, reduced 40%)

352-s I N O V E M B E R 1 9 7 6

tained at the center-thickness may be misleading.

3. While tearing is commonly associated with the elongated non-metallic inclusions, initiation may also occur by splitting along sheets of fine inclusions, textured ferrite bands, and embrittled grain boundaries.

4. Susceptibility tends to increase with an increase in the oxygen con­tent of the steel.

5. Larger inclusions tend to form voids earlier than the smaller ones.

6. It is unlikely that the inclusion count, using the prevailing tech­niques, would yield a satisfactory correlation with the lamellar tearing susceptibility, as these techniques disregard the smaller inclusions in the plate, which may exert significant influence on the tearing suscep­tibility, and the anisotropic effects in­duced by crystallographic texture, banding, and other microstructural factors.

Bibliography

1. "Lamel lar Tearing of Steel Worr ies Designers," Engineering News Record, Sept. 21 , 1972, p. 53.

2. "Lamel lar Tearing Subject of Law­suit," Ibid, March 20, 1975, pp. 39-40.

3. "AISC, A rmco named th i rd party defendants in lamellar tearing suit," Ib id, July 10, 1975, p. 16.

4 . " S t e e l c r a c k i n g c a u s e s d e s i g n change, delays j ob , " Ib id, January 29, 1976, p. 12.

5. Burdekin, F. M., "Lamel lar Tearing in Bridge Girders — A Case History," "Metal Construction and British Welding Journal, Vol. 3, No. 5, May 1971, pp. 205-209.

6. Goodger, A. H., "Fissuring along the Flow Structure of a Plate Under Fillet Welds," BSI News, Sept. 1966, pp. 10-13.

7. Watanabe, Masaki, "The Pull-Cut Fracture in Rolled Steel Plate," Sympo­sium, Welding in Shipbuilding, 1961, pp . 219-225, Institute of Weld ing, London.

8. Takeshi, Y., "Lamel lar Tearing and Marine Structures," Welding and Metal Fabrication, Vol. 43, No. 10, December 1975, pp. 740-746.

9. Meyer, H. J. , "Sub Surface Cracks in Plates at the Points of Appl icat ion of Stresses Perpendicular to the Plate Sur­face," Archiv Eisenhuttenw, 35 (9), 1964, pp. 903-908. (Translated BWRA Bulletin (6), Oct. 1965).

10. Wormington, H., "Lamel lar Tear­ing in Si l icon-Ki l led Boiler Plate," Welding and Metal Fabrication, Vol. 30, No. 9, Sept. 1962, p. 370-373.

11 . Nucleonics Week, Wor ld News, 8, Apr i l 17, 1969.

12. "Commentary on highly restrained welded connect ions," Engineering Jour­nal of AISC, 1973, pp. 61-73.

13. Thornton, C. H., "Quali ty Control in Design and Supervis ion Can Eliminate Lamellar Tear ing," Engineering Journal of AISC. Vol. 9, No. 6, 1973, pp. 112-116.

14. "Report on Enquiry into the Failure of Structural Members Stressed in the Direction of Plate Thickness," Welding in the World, Vol, 11 , No. 7/8, 1973, pp. 44 -47.

A i cty^.&£* \SUMf»

33!' W IW

ooox 4 0 0 0 X

Fig. 26 — Extent of void growth around inclusions in various steels: (a) semikilled A-285 (M); (b) semikilled A-36 (X); (c) Al killed A533 Gr.B (D); (d) Pb-sulfurized F. M. steel (N). (Reduced 35%)

*m

zS-j^14&X* lLrJ*x~ 5-tW • K V

OF c *

•v^eS* 4> "SOJjrp

Fig. 27 — Types of terrace linkage: (a) necking; (b) microvoid sheets; (c) zig-zag tearing; (d) intergranular cracking

WELDING RESEARCH S U P P L E M E N T I 353-s

! S T J - : I - 7 s-zry& Srt .

- I * ^L-*^,?*,- .>* ~» % -S? '

F/g. 28 — Types of iro/'d linkage: (a) necking; (b) microvoid coalescence; (c) quasi-cleavage; (d) cleavage. (Reduced 54%)

fig. 30 — Shear linkage of terrace tears: (a) microvoids on either side of shear wall; (b) microvoids on the terrace as well as shear wall. (Reduced 32%)

15. "Discussion on Lamellar Tear ing," Metal Construction and British Welding Journal, Vol. 1, No. 2, Feb. 1969, pp. 117-124.

16. Nagel , D. and Schonher r , W., "Strength and Deformation Propert ies of Structural Steels in the Direction of Thick­ness," Metal Construction and British Welding Journal, Vol. 1, No. 2, Feb. 1969,

pp. 64-67. 17. Heuschkel J., "Lamel lar Tear ing,"

Private Communicat ion with K. H. Koop­man, Feb. 9, 1971.

18. Vrbensky. J., "Some Weldabil i ty Problems of Thin Mn-V-N Type Stress Plates f rom the Crackabi l i ty Point of View," Metal Construction and British Welding Journal, Vol. 1. No. 2, Feb. 1969.

Fig. 29 — Lamellar tear and cleavage: (a) lamellar tear initiation and cleavage propagation; (b) transition zone from lamellar tear to cleavage; (c) sulfide within a void surrounded by cleavage facets; (d) cluster of dimples surrounded by cleavage. (Reduced 54%)

pp. 44-49. 19. Lombard in i , J. , "Crack ing as a

Criteria for Weldabi l i ty," Metal Construc­tion and BWJ, Vol. 1. No. 2, Feb. 1969, pp. 40-43.

20. Nicholls, D. M., "Lamel lar Tearing in Hot Rolled Steel," British Welding Jour­nal, Vol. 15. No. 3, March 1968, pp. 103-112.

2 1 . Amemya, Y., Satake, M., Kaj imoto, K. and Ohba. K.. "Effects of Welding Pro­cedure and Steel Materials on Prevention

• '•" i ^ »»"-" • " ot Lame lar Tear ing," International Insti­tute of Welding, Doc. No. IX-780-72, May 1972.

22. Oates, R. P. and Stout, R. D., "A Quantitative Weldabil i ty Test for Suscep­tibility to Lamellar Tear ing," Welding Jour­nal, Vol. 52, No. 11, Nov. 1973, pp. 4 8 1 -491.

23. Nishio, Y.. Yamamoto. Y., Kaj imoto, K. and Hirozane. T., "On Lamellar Tear­ing in Mult i run Fillet Welds," Technical Review, Vol. 9. No. 3. Oct. 1972, Mitsu­bishi Heavy Industries Ltd., Japan.

24. Bellen, A. Sparraft. M. J. and Van-der Veen, J . H., "Some Steelmakers ' Ex­perience on Improving the Resistance of Steel Plate to Lamellar Tear ing," First International Symposium of the Japan Welding Society, No. 8, 1971.

25. Elliott, D. N., "A Fractographical Examination of Lamellar Tearing in Mult i -run Fillet Welds," Metal Construction and British Welding Journal, Vol. 1, No. 2, Feb. 1969. pp. 50-57.

26. Farrar. J. C. M., Dolby, R. E. and Baker, R. G.. "Lamel lar Tearing in Welded Structural Steels," Welding Journal, 48 (7), July 1969, Research Suppl . , pp. 274s-282s.

27. Jubb, J. E. M., Curr ick, L. and Ham­mond, J., "Some Variables in Lamellar Tear ing." Metal Construction and British Welding Journal. Vol. 1, No. 2, Feb. 1969.

354 -s I N O V E M B E R 1 9 7 6

pp. 58-63. 28. Arita, Y. and Kaj imoto, K., "The

Study of Lamellar Tearing in Offshore Structure," Fourth Annual Offshore Tech­nology Conference in Houston, Texas, May 1972.

29. Owen, W. S., Cohen, M. and Aver-bach, B. L., "The Influence of Ferrite Banding on the Impact Propert ies of Mild Steel," Welding Journal, 37 (8), Aug. 1958, Research Suppl . , pp. 368s-374s.

30. Farrar, J . C. M. and Dolby, R. E., "An Investigation into Lamellar Tearing, Metal Construction and British Welding Journal, Vol. 1, No. 2, Feb. 1969, pp. 32-34.

3 1 . Arak i , M., et al., "Crack ing in Mu l -t irun Fillet Welds," Nippon Kokan Tech­nical Report-overseas, June 1974, pp. 25-35.

32. Nicholls, D. M., Wi lson, D. M. and J u b b , J . E. M., " C o r r e s p o n d e n c e on Lamellar Tear ing," British Welding Jour­nal, Vol. 13, No. 5, May 1966, pp. 326.

33. Gol ikov, I. N., "Certain peculiar­ities in the format ion of oxide- inclusions revealed by complex analysis," in 'Pro­duct ion and Appl icat ion of Clean Steels,' ISI, Sp. report , ISI London, 1970, pp. 35-4 1 .

34. Keissl ing, R., "The Influence of Non-Metal l ic Inclusions on the Propert ies of Steel," J. Metals, Vol. 22, No. 10, Oct. 1969, pp. 48-54.

35. Little, J. H. and Henderson, W. J . M., "Effect of Sulf ide Inclusions on the A n i s o t r o p y of Duc t i l e C h a r p y Shel f Energy," in 'Effect of Second Phase Par­ticles on the Mechanical Propert ies of Steel, ' ISI, Sp. Report, ISI, London, 1971, pp. 171-181.

36. Baker T. J. and Charles, J. A., "Morpho logy of Manganese Sulf ide in Steel," JISI, September 1972, pp. 702-6.

37. Baker, T. J. and Charles, J. A., "Deformat ion of Manganese Sulf ide In­clusions in Steel," Ibid, pp. 680-690.

38. Santen, S. and Ohman, B., "Tech­niques for obtaining extremely low con ­tents of sulfur in steel under oxidizing and reducing condi t ions," Clean Steel, v. 1, Report of the Royal Swedish Academy of Engineering Sciences, 169:1, 1971, pp. 7-19.

39. Pickering, F. B., "Effect of Pro­cessing Parameters on the Origin of Non-metallic Inclusions, in "Effect of Second Phase . . .," ISI, Sp. Report, ISI, London, 1970, pp. 75-91 .

40. Keissling, R. and Nordberg , H., " I n ­

f luence of Inclusions on the Mechanical Propert ies of Steel" in 'Product ion and App l i ca t i on of C lean Steels ' ISI, Sp . Report, ISI, London, 1970, pp. 179-185.

4 1 . Brooksbank, D. and Andrews, K. W., "Stress Fields Around Inclusions and Their Relation to Mechanical Propert ies," Ibid, pp. 186-198.

42. A l lmand, T. R. and Blank, J. R., "An Evaluation of the Quant imet for Assess­ing Non-Metal l ic Inclusions and Other Microscopical Features in Metals and Alloys," in 'Automatic Cleanness Assess­ment of Steel , " ISI, Sp. Report, No. 112, ISI, London, 1968, pp. 47-60.

43. Mel ford, "Design and Use of Auto­matic Instruments for Cleanness Assess­ment," Ibid., pp. 14-23.

44. Charpent ier, P. L. et al., "Black- l ine s u r f a c e d e f e c t s in a l u m i n u m - k i l l e d steels," Metals Technology, Vol . 2, part 2, February 1975, pp. 52 -61 .

45. Cairns, R. L. and Charles, J . A., "Banding in carbon and low-alloy steel — a brief review," Iron and Steel, Vol. 39, November 1966, pp. 511-515.

46. Jatczak, E. F., Girard i , D. J . and Rowland, E. S., "On Banding in Steel," Trans. A.S.M., Vol. 48, 1956, pp. 279-305.

47. Grange, R. A., "Effect of Micro-structural Banding in Steel," Metallurgical Transactions, Vol. 2, February 1971, pp. 417-426.

48. Coleman, T., Dulien, D. and Gouch, A., "The structure and propert ies of con ­trol led processed high strength ferr i te-pearlite steels," in Microstructure and Design of Alloys, Vol. 1, 1974, London, The Metals Society, pp. 70-74.

49. Hero, H., Evensen, J. and Embury, J . D., "The occurrence of delaminat ion in a control led rol led HSLA steel," Canadian Met. Quarterly, Vol. 14, No. 2, 1975, pp. 117-122.

50. Yen, C. M. and Stickels, C. A., "Lamel late fracture in 5150 steel pro­cessed by modif ied ausforming," Metal­lurgical Transactions, Vol. 1, Nov. 1970, pp. 3037-3047.

51 . Stoloff, N. S., "Effects of alloying on fracture character ist ics," Fracture, Vol. VI, Edited by Liebowitz, H., Academic Press, N.Y. 1969, pp. 1-81.

52. Boniszewski, T. and Moreton, Mrs. J . , "Effect of microvoids and manganese sulf ide inclusions in steel on hydrogen evo lu t ion and embr i t t l emen t , " British Welding Journal, Vol. 14, No. 6, June 1967, pp. 321-336.

53. Moreton, Mrs. J. , Coe, F. R. and

Boniszewski, T., "Hydrogen movement in weld metals," Metal Construction and British Welding Journal, Vol. 3, No. 5, May 1971, pp. 185-187.

54. Hewitt, J . and Murray, J. D., "Effect of sulfur on the product ion and fabr ica­tion of C-Mn steel forg ings," Ibid, Vol. 1, No. 2, Feb. 1969, pp. 24 -31 .

55. Smith, N. and Bagnall , B. I., "The influence of sulfur on HAZ cracking of C-Mn steel welds," Ibid, pp. 17-23.

56. "Discussion of hydrogen induced cold crack ing, " Ibid. pp. 109-116.

57. Argon, A. S. and Im, J . , "Separa­t ion of Second Phase Particles in Sphero-dized 1045 Steel, Cu-0.6% Cr Alloy and Marag ing Steel in Plast ic S t ra in ing , " Metallurgical Transactions A, Vol. 6A, Apri l 1976, pp. 839-851.

58. Gangulee, A. and Gur land, J., "On the Fracture of Sil icon Particles in A lu ­minum-Si l icon Al loys," TMS-AIME, Vol. 239, Feb. 1967, pp. 269-272.

59. Gur land, J. , "Observat ions on the Fracture of Cemen t i t e Par t ic les in a Spherodized 1.05% C Steel Deformed at Room Temperature," Acta Met., Vol. 20, May 1972, pp. 735-74.

60. Cox, T. B. and Low, J. R., "An Inves­tigation of the Plastic Fracture of AISI 4340 and 1 8 - N i c k e l - 2 0 0 G r a d e M a r a g i n g Steels," Metallurgical Transactions, Vol. 5, June 1974, pp. 1457-1470.

6 1 . Van Stone, R. H. and Psioda, J. A., "D iscuss ion of Meta l lu rg ica l Factors Affecting Fracture Toughness of A lu ­minum Al loys," Metallurgical Transac­tions A, Vol. 6A, Apri l 1975, pp. 668-670.

62. Argon, A. S., Im, J. and Safoglu. R., "Cavity Formation f rom Inclusions in Duc­tile Fracture," Metallurgical Transactions A, Vol. 6A, Apr i l 1975, pp. 825-836.

63. Schulz, B. J. and McMahon, C. J., Jr., "Al loy Effects in Temper Embrit t le­ment" in "Temper Embri t t lement of Alloy Steels," ASTM STP 499, ASTM. 1972, pp. 104-135.

64. Beachem, C. D. and Yoder, G. R., " E l a s t i c - P l a s t i c F r a c t u r e by H o m o ­geneous Microvoid Coalescence Tearing Along Alternating Shear Planes," Met. Trans., Vol. 4, Apri l 1973, pp. 1145-1153.

65. Chipperf ie ld, C. G. and Knott, J. F., "Microst ructure and Toughness of Struc­tural Steels," Metals Technology, Vol. 2, Feb. 1975, pp. 45-51 .

66. Knott, J. F., "Mechanics and Mech­anisms of Large Scale Britt le Fracture in Structural Metals," Materials Science and Engineering, Vol 7, Jan. 1971, pp. 1-36.

WRC Bulletin 216 June 1976 "P reven t ing Hydrogen- Induced Cracking Af te r W e l d i n g of Pressure Vessel Steels by Use of Low T e m ­

perature Pos twe ld Heat T r e a t m e n t s " by J . S. Caplan and E. Landerman

Hydrogen-induced cracking occurs either in the heat-affected zone microstructure or in weld metal when four factors react simultaneously. These factors have been defined as (1) presence of hydrogen. (2) welding stresses, (3) a susceptible microstructure and (4) a low temperature. Hydrogen can become available during welding from base and welding materials and extraneous contaminating matter. Data are presented to show the effects of preheat and postweld heat treatments. These data are principally concerned with the type of steels used for nuclear pressure vessels.

Publication of this paper was sponsored by the Pressure Vessel Research Committee of the Weld­ing Research Council.

The price of WRC Bulletin 216 is $6.50 per copy. Order should be sent with payment to the Weld­ing Research Council, United Engineering Center, 345 East 47th Street, New York, N.Y. 10017.

W E L D I N G R E S E A R C H S U P P L E M E N T ! 3 5 5 - s